This patch set expands the set of dedicated RGB->YUV unscaled functions to help with encoding camera output on a Pi. Obviously there are other uses but that was the motivation.
It enforces the general bitexact path for the fate tests that depend on it. It renames the existing bgr function as bgr... so we don't end up with the counterintuative situation where BGR is handled by rgb... and BGR would be handled by rgb.. Adds RGB functions Improves the rounding in the dedicated function as that improves its score when tested with test/swscale and fixes it to allow any width (contrary to the comment any height was already allowed). Adds XRGB->YUV functions to complete the set Adds Aarch64 neon for BGR24 & RGB24 I haven't built fate tests for this as I'm not quite sure what the appropriate tests would be. The x86 asm doesn't match either the C template with improved rounding or the previous template (I'm not quite sure what it does but it produces a different score out of tests/swscale to either method) so a simple results match isn't going to work. Regards John Cox John Cox (6): fate-filter-fps: Set swscale bitexact for tests that do conversions swscale: Rename BGR24->YUV conversion functions as bgr... swscale: Add explicit rgb24->yv12 conversion swscale: RGB24->YUV allow odd widths & improve C rounding swscale: Add unscaled XRGB->YUV420P functions swscale: Add aarch64 functions for RGB24->YUV420P libswscale/aarch64/rgb2rgb.c | 8 + libswscale/aarch64/rgb2rgb_neon.S | 356 ++++++++++++++++++++++++++++++ libswscale/bayer_template.c | 2 +- libswscale/rgb2rgb.c | 25 +++ libswscale/rgb2rgb.h | 23 ++ libswscale/rgb2rgb_template.c | 174 +++++++++++++-- libswscale/swscale_unscaled.c | 114 +++++++++- libswscale/x86/rgb2rgb_template.c | 13 +- tests/fate/filter-video.mak | 4 +- 9 files changed, 694 insertions(+), 25 deletions(-) -- 2.39.2 _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".